First thing first, we need to use the X-ray images and for this purpose I used the repo of Dr. Cohen(He started to collect X-ray images and he is pushing these images in his github repo) https://github.com/ieee8023/covid-chestxray-dataset
In order to create the COVID-19 X-ray image dataset for this project, I:
Parsed the metadata.csv
file found in Dr. Cohen’s repository.
Selected all rows that are:
Positive for COVID-19 (i.e., ignoring MERS, SARS, and ARDS cases).
Posterioranterior (PA) view of the lungs. I used the PA view as, to my knowledge, that was the view used for my “healthy” cases, as discussed below; however, I’m sure that a medical professional will be able clarify and correct me if I am incorrect (which I very well may be, this is just an example).
In total, that left me with 25 X-ray images of positive COVID-19 cases (Figure 2, left).
The next step was to sample X-ray images of healthy patients.
To do so, I used Kaggle’s Chest X-Ray Images (Pneumonia) dataset (https://www.kaggle.com/paultimothymooney/chest-xray-pneumonia) and sampled 25 X-ray images from healthy patients (Figure 2, right). There are a number of problems with Kaggle’s Chest X-Ray dataset, namely noisy/incorrect labels, but it served as a good enough starting point for this proof of concept COVID-19 detector.
After gathering my dataset, I was left with 50 total images, equally split with 25 images of COVID-19 positive X-rays and 25 images of healthy patient X-rays.
I’ve included my sample dataset in the “Downloads” section of this tutorial, so you do not have to recreate it.
The main python code for this project is train_covid19.py. and for executing this code you need also mention the dataset for the execution, therefore it will be python3 train_covid19.py --dataset dataset, besides here is two other command line arguments which could be used: --plot : An optional path to an output training history plot. By default the plot is named plot.png unless otherwise specified via the command line. --model : The optional path to our output COVID-19 model; by default it will be named covid19.model . For this project I split the dataset in 80/20 which is 80% of the data for training and 20% for the testing. In order to ensure that our model generalizes, we perform data augmentation by setting the random image rotation setting to 15 degrees clockwise or counterclockwise.
Training our COVID-19 detector with Keras and TensorFlow With our train_covid19.py cript implemented, we are now ready to train our automatic COVID-19 detector over the dataset that we have. Once the train is done we could see the following results below: [INFO] evaluating network... precision recall f1-score support covid 0.83 1.00 0.91 5 normal 1.00 0.80 0.89 5 accuracy 0.90 10 macro avg 0.92 0.90 0.90 10 weighted avg 0.92 0.90 0.90 10
acc: 0.9000 sensitivity: 1.0000 specificity: 0.8000
As you can see from the results above, our automatic COVID-19 detector is obtaining ~90-92% accuracy on our sample dataset based solely on X-ray images — no other data, including geographical location, population density, etc. was used to train this model.
We are also obtaining 100% sensitivity and 80% specificity implying that:
- of pients that do have COVID-19 (i.e., true positives), we could accurately identify them as “COVID-19 positive” 100% of the time using our model.
- Of patients that do not have COVID-19 (i.e., true negatives), we could accurately identify them as “COVID-19 negative” only 80% of the time using our model.
As our training history plot shows, our network is not overfitting, despite having very limited training data:
And last by not least here is the Final result from the x-ray which shows either the x-ray image has the potential of COVID-19 or not.
